There is a new king in the world of games. Meet Student of Games.
A team of artificial intelligence researchers from Equilibrium Technologies, Sony AI, Amii, and Midjourney, working in conjunction with Google's DeepMind project, has created an artificial intelligence system called "Student of Games" (SoG). This system can not only beat a person in a variety of games, but also independently learn how to play new ones. The results of the study were published in the journal Science Advances.
For the past fifty years, computer scientists and engineers have been developing concepts for machine learning and artificial intelligence, using human-generated data to train computer systems. These technologies are used in a wide variety of scenarios, including board games and salon games.
Teaching a computer to play a board game and then improve its abilities to the point where it can beat a human was a milestone of sorts, demonstrating how far artificial intelligence has evolved. In this new study, the research team has taken another step toward creating artificial general intelligence, where a computer can perform tasks considered superhuman.
Until now, most computer systems designed for playing board games have focused on a single game, such as chess. Creating such systems, scientists have developed a certain type of limited artificial intelligence. In a new initiative, researchers have built an intelligent system that can play a variety of games that require different skills.
There are two main types of games in gameplay: those with complete and incomplete information. The first are games in which both players have complete information about the game, for example, about the location of all the pieces on the board. Others, such as poker, assume that only a fraction of the information is known to each player. SoG can play both types of games and beat human experts in them.
At the moment, the system has already beaten other AI systems and humans in games such as Go, chess, Scotland Yard, and Texas Hold'em, and the research team suggests that it is likely to succeed in other types of games, as it is able to learn to play almost any game on its own.
A team of artificial intelligence researchers from Equilibrium Technologies, Sony AI, Amii, and Midjourney, working in conjunction with Google's DeepMind project, has created an artificial intelligence system called "Student of Games" (SoG). This system can not only beat a person in a variety of games, but also independently learn how to play new ones. The results of the study were published in the journal Science Advances.
For the past fifty years, computer scientists and engineers have been developing concepts for machine learning and artificial intelligence, using human-generated data to train computer systems. These technologies are used in a wide variety of scenarios, including board games and salon games.
Teaching a computer to play a board game and then improve its abilities to the point where it can beat a human was a milestone of sorts, demonstrating how far artificial intelligence has evolved. In this new study, the research team has taken another step toward creating artificial general intelligence, where a computer can perform tasks considered superhuman.
Until now, most computer systems designed for playing board games have focused on a single game, such as chess. Creating such systems, scientists have developed a certain type of limited artificial intelligence. In a new initiative, researchers have built an intelligent system that can play a variety of games that require different skills.
There are two main types of games in gameplay: those with complete and incomplete information. The first are games in which both players have complete information about the game, for example, about the location of all the pieces on the board. Others, such as poker, assume that only a fraction of the information is known to each player. SoG can play both types of games and beat human experts in them.
At the moment, the system has already beaten other AI systems and humans in games such as Go, chess, Scotland Yard, and Texas Hold'em, and the research team suggests that it is likely to succeed in other types of games, as it is able to learn to play almost any game on its own.
